In Reply to: RE: Sonic and Other --No posted by fmak on November 27, 2011 at 07:56:24:
The link described what goes on in computer systems used with audio. It said nothing more or less than what many of us have said. I don't disagree with these facts. At issue is what to do about them.
My personal belief is that apart from a few simple things, messing around in a computer system is likely to be extremely frustrating and a great waste of time. There is little chance to "optimize" anything, since changes to any one part interact in an incomprehensibly complex fashion with other parts of these system. Furthermore, what tweaks work best with one DAC may not work so well with another. From a designer's perspective it is much simpler to solve these problems downstream from the computer in the DAC or in a separate "reclocker" box.
With the proper clock architecture and logic design it is possible to avoid logic triggering off of noisy input signals. (How to do this has been known since the 1960's. I have some dusty old books that describe how communications adapters deal with these problems.) Furthermore, when using clock architectures that are timed off of the input signals it is possible to isolate the portions that potentially have problems so that they can be logically, electrically and physically isolated from the remainder of a DAC. (If necessary, this can involve using an external reclocker.) In the ESS SABRE white paper the authors describe how they handle the clock input circuitry that they use with their SPDIF receiver circuitry. Basically, they have as little logic as possible running in the external clock domain and synchronize it with the local master clock. If the SPDIF source has been synchronized with the local master clock then there will be no need to use the on-chip ASRC. (More details of how this chip works can be inferred from reading the patents.)
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
This post is made possible by the generous support of people like you and our sponsors:
Follow Ups
- RE: Sonic and Other --No - Tony Lauck 10:11:44 11/27/11 (11)
- RE: Sonic and Other --No-You - fmak 20:12:00 11/27/11 (10)
- RE: Sonic and Other --No-You - Tony Lauck 07:12:18 11/28/11 (9)
- so - fmak 07:15:47 11/28/11 (8)
- RE: so - Tony Lauck 08:39:43 11/28/11 (7)
- RE: so - rick_m 09:11:17 11/28/11 (6)
- RE: so - Tony Lauck 12:26:28 11/28/11 (5)
- RE: so-Ground Loops - fmak 01:24:04 11/30/11 (0)
- RE: so - rick_m 20:34:34 11/28/11 (3)
- Drunken moose, measuring jitter rejection of DACs, and general quality of AES engineering - Tony Lauck 09:09:25 11/29/11 (2)
- RE: Drunken moose, measuring jitter rejection of DACs, and general quality of AES engineering - rick_m 23:02:11 11/29/11 (1)
- RE: Drunken moose, measuring jitter rejection of DACs, and general quality of AES engineering - Tony Lauck 07:43:47 11/30/11 (0)